1913 Phillip . Migration and Periodic Accuracy L95 



ESifrig'a observations at Ottawa, Canada, 1 show thai this excep- 

 tional ea "ii delayed some birds, like the Hummingbird, for two 

 weeks, while other species were verj little affected. Tin's, however, 

 only means thai the racial accuracy of the species was forcibly 

 interfered with from outside, in some cases far more seriously than 

 in others. 



We inn i, however, now consider another side of the question. 

 So far we have been dealing with 'firsl arrh als.' What we si ion Id 

 like, of course, would !><■ individual arrivals, but until we have some 

 definite information <»n this point, which can only be supplied by 

 the return <>f banded birds to nesting sites, we must contenl our- 

 selves with the consideration of the possibh errors in recording 

 lii i arrivals, and also with the highly interesting 'bulk arrivals,' 

 or dates of greatesl frequency of transienl migrants. 



The question «>f the error of firsl arrivals has been discussed by 

 Messrs. Stone and Cooke. Stone found thai manj eyes were 

 Letter than one, and Cooke 8 states thai where one observer wa 

 op< rating, liis dates of first arrival were apt to be over a day late 

 on an average. Cooke's method of averaging migration arrivals 4 

 consists in throwing out dates which are more than six daj out 

 of the way, hi, experience teaching him tliat "birds seldom vary 

 on accounl <>f the season more than six days either way from the 

 average date of their arrival." This method may seem to some, 

 .1 to the u liter, rather arbitrary. 



Mr. Stone gives a good discussion of the problem of recording 

 early arrn als in the Proceedingsof the Academj <>f Natural Science i 

 of Philadelphia ' for 1908. In it he shows thai togetal the actual 

 date we must combine the records of many observers at different 

 point a feu miles apart, and combine them in certain definite 

 ways, so as to throw oul stations in which individual error or in- 

 compatibility of surroundings delayed the detection of the pecie 

 in question. I5y ' l»ull< arrival ' Mr. Stone means the date on which 

 the 1 1. cie ha arrived at Imlf the stations in a given re tricted area. 

 He regard the hulk arrival, or greatesl frequencj as recorded only 



i Auk, 1008, p I 



torn u . Condor, 1006, p. hh. 

 Oooke w w . \ui. 1907, p 346. 

 < Oooke, w. w , Auk, L908, p 186 



